Low-Power Wide Area Networks (LP-WANs) are emerging as a promising solution for connecting Internet of Things and Machine Type Communication devices. If ultra-narrowband (UNB) networks, a subclass of LP-WANs, reach predicted deployment numbers and densities, they will face two challenges: inter-technology and intra-technology interference. This paper proposes the first experimental architecture designed for the optimization of UNB networks. We illustrate its implementation on a case study of a SIGFOX network and the resulting extension of the existing LOG-a-TEC testbed. The proposed architecture enables context data collection, context model development, optimization and transmission control using rapid experimentation cycle approach enabled by flow based programming using Node-RED. Through preliminary results, we show the feasibility of PHY and MAC context data collection, point out challenges that are specific for UNB context modeling and discuss options for optimization. All datasets, context modeling and optimization tools used in the paper will be released as open source.

This paper has been presented at the Computer and Networking Experimental Research Using Testbeds, IEEE Infocomm 2017. The link to the paper: https://www.researchgate.net/publication/316605730_Optimization_of_Ultra-Narrowband_Wireless_Communication_an_Experimental_Case_Study